Title: You Can't Hide Behind Your Headset: User Profiling in Augmented and Virtual Reality

Abstract: Virtual and Augmented Reality (VR, AR) are increasingly gaining traction thanks to their technical advancement and the need for remote connections, recently accentuated by the pandemic. Remote surgery, telerobotics, and virtual offices are only some examples of their successes. As users interact with VR/AR, they generate extensive behavioral data usually leveraged for measuring human behavior. However, little is known about how this data can be used for other purposes.
In this work, we demonstrate the feasibility of user profiling in two different use-cases of virtual technologies: AR everyday application ($N=34$) and VR robot teleoperation ($N=35$). Specifically, we leverage machine learning to identify users and infer their individual attributes (i.e., age, gender). By monitoring users' head, controller, and eye movements, we investigate the ease of profiling on several tasks (e.g., walking, looking, typing) under different mental loads. Our contribution gives significant insights into user profiling in virtual environments.